Intimate partner violence (IPV), including physical and psychological aggression, is associated with numerous negative outcomes (e.g., depression, poor physical health) that significantly impair the functioning of individuals and their families and have high costs for society at large. For both men and women, physical IPV victimization was associated with increased risk of current poor health, depressive symptoms, substance use, developing a chronic disease, chronic mental illness, and injury [1]. In a related population-based study with 4,060 women, Hispanic women in the United States who were IPV victims had disproportionate rates of depression, post-traumatic stress disorder, substance misuse, and anxiety, in decreasing orders of prevalence, i.e., with depression having the highest prevalence [2].
Despite the research on psychological characteristics of perpetrators of IPV, research into the biological factors that underlie the perpetration of IPV has been limited. With the exception of brain imaging for highly aggressive men [3], the field of aggression research has largely relied on self-report, observer-report, observation of dyads, and interview data in order to identify risk factors (e.g., personality traits, attitudes towards aggression, and exposure to aggression in the family or origin) [3, 4]. Recently, organizations such as the National Institute of Mental Health have emphasized the importance of understanding phenomena through direct observable behavior as well as through neurobiological measures [5]. Thus, biomarkers, or biological states that differentiate between aggressive and non-aggressive individuals, represent an important and necessary next step in the evolution of aggression research.
The field has reasonably complex multivariate psychosocial models of intimate partner aggression [6], but Raine [7] has argued for evaluations of both psychological as well as biological predictors of aggression. Accordingly, we turn to discussion of biological markers of aggression, along with analyses of how such markers might be used in an interactive or additive manner with psychosocial risk factors. Further, if one can use biological markers of aggression that can be obtained readily in an office, such as salivary hormonal assays, the practical utility of such markers increases. A literature review of biological markers of IPV perpetration [8] suggests that biological variables in the domains of head injury, neuropsychology, psychophysiology, neurochemistry, metabolism and endocrinology, and genetics play a significant role in the etiology of IPV. The authors concluded that at the most basic level, neurochemical alterations in perpetrators, specifically excessive testosterone or reduced serotonin activity, reflect an alteration of neuronal function that can be simplistically thought of as promoting rapid responding to threatening external stimuli. An analysis of archival data from 4,462 U.S. military veterans concluded that testosterone leads directly to antisocial behavior since testosterone was correlated with a variety of antisocial behaviors for all individuals [9]. However, a review of over 42 correlational studies concluded that there is a small association (r = 0.08) between testosterone levels and measures of aggression and that these associations were strongest for young men and offenders [10]. Archer [11] also concluded that people with higher existing levels of testosterone are more likely to show higher scores on a variety of different assessments of dominance, although this is a weak relationship. The relatively weak association between testosterone and aggression has led some to argue that this weak association may be due to the failure to account for levels of a second hormone, namely, cortisol. Thus, the Dual-Hormone Hypothesis proposes that the hormones testosterone and cortisol jointly regulate social dominance and aggression in humans. The neuroendocrine systems that produce testosterone and cortisol are thought to be diametrically opposed, with cortisol modulating the effects of testosterone on aggressive psychopathology [12, 13]. The combination of high levels of testosterone (associated with dominance-seeking behavior), and low levels of cortisol (associated with avoidance behavior) may be associated with increased dominance and aggression more consistently than the levels of testosterone or cortisol, individually [13].
A few studies have investigated the role of testosterone and cortisol on intimate partner violence. Romero-Martinez et al. [14] compared participants who had previously been jailed for IPV and controls matched for SES and absence of partner aggression on testosterone and cortisol levels. Their methodology involved having subjects stressed by performing the Trier Social Stress Test (TSST). IPV perpetrators experienced decreases in salivary testosterone (T) levels, a moderate worsening of mood, slight anxiety, and a salivary cortisol (C) level increase. Moreover, high basal T was related to high levels of anger, anxiety, and worse mood. Controls experienced smaller changes in T and larger changes in C and psychological mood. The authors concluded that together with social aspects involved in IPV, differences in psychobiological variables and their relationships could play a relevant role in the onset and perpetuation of violent behavior. In a follow-up study by the same lab group, Romero-Martinez et al. [15] compared IPV perpetrators with men matched for SES and no IPV. They found that while IPV perpetrators had higher antisocial, borderline, and narcissistic personality traits and anger expression than controls, they did not differ in basal T/C ratio. However, only in IPV perpetrators was there a positive relationship between these variables, the T/C ratio playing a moderating role in the relationship of antisocial and borderline traits with anger expression. This led the researchers to conclude that in IPV perpetrators the T/C ratio may explain why certain personality traits are associated with high risk of becoming violent. In a third study from the same lab group in Valencia, Spain, IPV perpetrators were compared with men with no IPV who were matched for age and SES [16]. Both the IPV perpetrators and the non IPV perpetrators were stressed using the Trier Social Stress Test. Perpetrators of IPV against women had lower salivary cortisol and higher salivary testosterone/cortisol ratio levels during the post- acute cognitive laboratory stressor period, as well as higher total levels (average) of salivary oxytocin than controls. In addition, high levels of baseline anxiety and negative affect were related to high rises in cortisol during the stress task only in the perpetrators.
An additional study from a different lab group evaluated the association of T/C with IPV of male undergraduate college students using the Trier Social Stress Task [17]. Trait aggression moderated the relationship between the ratio of testosterone to cortisol (T/C) and IPV perpetration. High T/C ratio, or more testosterone relative to cortisol, was associated with elevated IPV in men low in trait aggression, whereas the association between T/C ratio and IPV was non-significant in men high in trait aggression [17]. This result seems counter to the overall result of others -where Archer [10] found the association greatest in young men and criminals.
Overall, a major review by Archer et al. [10] concluded that the correlation between T and aggression is small (-.08). In addition, the association, if detected, appeared to be evident in young men and criminal offenders. Later research looked at the simultaneous role of T and C as it was believed that C would moderate or dampen the role of T. Research on the simultaneous role of T and C has proven to be complex as it is unclear if differences in T would be evident as a baseline measure in men or whether men need to be stressed to detect differences in T as some believe that rises when males are threatened. Similarly, should differences in the stress hormone, cortisol, across groups of aggressive and non-aggressive men be evident during a baseline non-stress period or only when the men are stressed, and cortisol has risen. Recent research in the T/IPV arena has used men who have been charged with physical assault against their female partners to obtain men who engage in more severe levels of aggression, T and C are evaluated simultaneously, and some stressor is used to potentially elevate both T and C.
The Present Study
In order to gain a better understanding of the individual and joint impact of testosterone and cortisol on behavior, the present study was developed to test the differences in each hormone alone and conjointly between perpetrators of IPV on probation and non-violent controls. Differences in cortisol following exposure to a stressful event were also examined. In addition, this study assessed the moderation of the testosterone hypothesis and the dual hormone hypothesis by trait anger, using measures of testosterone and cortisol for the prediction of IPV.
Procedures included two laboratory visits consisting of questionnaires, and a number of salivary testosterone and cortisol collections. First, we hypothesized that perpetrators of IPV would have higher baseline testosterone, lower cortisol than the non-aggressive controls and that perpetrators would have a greater increase from pre to post stressor cortisol. Second, we hypothesized that a ratio of high testosterone to cortisol (T/C) would differentiate between male perpetrators of IPV and those men in the control group who have no history of aggression toward their partner. A third prediction was that trait anger would moderate the relationship between testosterone and perpetration of IPV. In addition, the current study proposed to test the Hypothesis 4) moderation of the dual hormone hypothesis by trait anger, using measures of testosterone and cortisol for the prediction of IPV.